--- Log opened Mon Jan 27 00:00:12 2014 | ||
-!- Krishna_ [80689984@gateway/web/freenode/ip.128.104.153.132] has joined #shogun | 00:34 | |
Krishna_ | I recently installed shogun 3.0.0 | 00:36 |
---|---|---|
Krishna_ | The install process went smoothly | 00:36 |
Krishna_ | but I can't locate the .so or the .mex files | 00:37 |
Krishna_ | for the interface install | 00:37 |
-!- Krishna_ [80689984@gateway/web/freenode/ip.128.104.153.132] has quit [Ping timeout: 245 seconds] | 04:28 | |
-!- lisitsyn [~lisitsyn@31-28-48-250.clients.tlt.100megabit.ru] has quit [Quit: Leaving.] | 05:48 | |
-!- tejasnikumbh [67157f4e@gateway/web/freenode/ip.103.21.127.78] has joined #shogun | 06:30 | |
tejasnikumbh | Hey guys | 06:30 |
tejasnikumbh | Are any of the gsoc metors here | 06:30 |
tejasnikumbh | I kinda have some ideas related to gsoc for this year which could be incorporated in shogun | 06:30 |
tejasnikumbh | I'm posting them on the mailing list. | 06:31 |
tejasnikumbh | Thanks | 06:31 |
tejasnikumbh | in the replies | 06:31 |
tejasnikumbh | Please let me know about your availablity on irc | 06:31 |
tejasnikumbh | :) | 06:31 |
-!- tejasnikumbh [67157f4e@gateway/web/freenode/ip.103.21.127.78] has quit [Quit: Page closed] | 06:37 | |
-!- tejasnikumbh [67157c4e@gateway/web/freenode/ip.103.21.124.78] has joined #shogun | 08:52 | |
-!- tejasnikumbh [67157c4e@gateway/web/freenode/ip.103.21.124.78] has quit [Ping timeout: 245 seconds] | 09:02 | |
-!- new_lido [~walid@41.218.173.186] has quit [Quit: Leaving] | 10:44 | |
-!- benibadman [~benibadma@94.135.236.129] has joined #shogun | 11:48 | |
-!- tejasnikumbh [67157c4e@gateway/web/freenode/ip.103.21.124.78] has joined #shogun | 12:01 | |
-!- tejasnikumbh [67157c4e@gateway/web/freenode/ip.103.21.124.78] has quit [Quit: Page closed] | 12:06 | |
-!- parijat [671b082a@gateway/web/freenode/ip.103.27.8.42] has joined #shogun | 12:12 | |
-!- parijat [671b082a@gateway/web/freenode/ip.103.27.8.42] has quit [Client Quit] | 12:16 | |
-!- lisitsyn [~lisitsin@mxs.kg.ru] has joined #shogun | 12:58 | |
-!- lisitsyn1 [~lisitsin@mxs.kg.ru] has quit [Ping timeout: 260 seconds] | 12:59 | |
-!- lisitsyn [~lisitsin@mxs.kg.ru] has quit [Read error: Operation timed out] | 13:04 | |
-!- lisitsyn [~lisitsin@mxs.kg.ru] has joined #shogun | 13:05 | |
-!- tom__ [2eda6d58@gateway/web/freenode/ip.46.218.109.88] has joined #shogun | 13:27 | |
tom__ | Hi all | 13:27 |
@wiking | hey | 13:27 |
tom__ | A question about: HistogramIntersection Kernel | 13:28 |
tom__ | What type of features can I use with that ? | 13:28 |
tom__ | HashDocDot feats are derived from dot feats, does it match ? | 13:28 |
tom__ | HistogramIntersection should use dense features, is there any way to convert string feats, (or hashdocdot feats) into dense feats ? | 13:29 |
@wiking | tom__: mmm well maybe it'd be better to convert HistogramIntersectionKernel to support other features | 13:37 |
tom__ | When StringFeatures refers to occurence histograms (of ngram etc ..), is it possbile to convert to dense features ? | 13:43 |
@wiking | mmm | 13:44 |
@wiking | haven't been working with stringFeats ... | 13:44 |
@wiking | maybe somebody else would have an opinion on this...? | 13:45 |
-!- tom__ [2eda6d58@gateway/web/freenode/ip.46.218.109.88] has quit [Ping timeout: 245 seconds] | 14:14 | |
-!- lisitsyn [~lisitsin@mxs.kg.ru] has quit [Quit: Leaving.] | 16:49 | |
-!- parijat [671b082a@gateway/web/freenode/ip.103.27.8.42] has joined #shogun | 17:37 | |
-!- benibadman [~benibadma@94.135.236.129] has quit [Ping timeout: 272 seconds] | 17:44 | |
-!- parijat [671b082a@gateway/web/freenode/ip.103.27.8.42] has left #shogun [] | 18:01 | |
-!- lisitsyn [~lisitsyn@80.252.20.67] has joined #shogun | 18:18 | |
-!- [paladin] [73f91219@gateway/web/freenode/ip.115.249.18.25] has joined #shogun | 18:27 | |
[paladin] | Hey. I'm interested in discussing about participating in GSoC 2014. Who do I get in touch with ? | 18:29 |
lisitsyn | [paladin]: hey | 18:46 |
[paladin] | hey. | 18:47 |
lisitsyn | I can talk about it I think :) | 18:47 |
lisitsyn | is there anything you want to ask? | 18:47 |
[paladin] | Yeah. | 18:48 |
[paladin] | One. I haven't seen any official ideas page for GSoC 2014. So I was wondering what you guys had planned ? | 18:49 |
lisitsyn | [paladin]: we are working on possible ideas these days | 18:50 |
[paladin] | Two: do you guys have benchmarking methods implemented ? | 18:50 |
lisitsyn | [paladin]: examples of ideas are GPs, random forests, MCMC and some frameworks | 18:51 |
lisitsyn | [paladin]: no not really | 18:51 |
lisitsyn | we had one in dotfeatures iirc | 18:52 |
[paladin] | What about an automatic benchmarking tool ? | 18:52 |
lisitsyn | but not as a general practice | 18:52 |
lisitsyn | if you want to propose such an idea it would be welcome if we are accepted :) | 18:52 |
lisitsyn | however we need to think what is it and how should it be done | 18:53 |
-!- HeikoS [~heiko@nat-164-138.internal.eduroam.ucl.ac.uk] has joined #shogun | 18:53 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 18:53 | |
[paladin] | Okay so should I approach when you have a page out to officially discuss ? Later ? | 18:56 |
lisitsyn | [paladin]: well some kind of | 18:57 |
lisitsyn | although it would be cool if you propose your idea in the mailing list | 18:57 |
lisitsyn | so we can comment etc | 18:58 |
[paladin] | okay. I think that would give me more time to sort of formalise it | 18:58 |
lisitsyn | [paladin]: sure! | 19:00 |
[paladin] | great. thanks. | 19:00 |
-!- [paladin] [73f91219@gateway/web/freenode/ip.115.249.18.25] has quit [] | 19:01 | |
-!- [paladin] [73f91219@gateway/web/freenode/ip.115.249.18.25] has joined #shogun | 19:02 | |
-!- [paladin] [73f91219@gateway/web/freenode/ip.115.249.18.25] has left #shogun [] | 19:03 | |
-!- killer_ [671b082a@gateway/web/freenode/ip.103.27.8.42] has joined #shogun | 20:01 | |
-!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has joined #shogun | 20:02 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 20:02 | |
-!- zxtx [~zv@c-98-223-196-32.hsd1.in.comcast.net] has quit [Ping timeout: 248 seconds] | 20:15 | |
-!- killer_ [671b082a@gateway/web/freenode/ip.103.27.8.42] has quit [Quit: Page closed] | 20:15 | |
-!- ChanServ changed the topic of #shogun to: Shogun Machine Learning Toolbox | Next Stammtisch Monday 2014-02-03 19-24hrs UTC |Devs are usually around between 10am-10pm UTC | When asking questions please be patient or use the mailing list. We always answer but it might take a while. | Channel logs http://bit.ly/1eBecWl | 20:24 | |
@sonney2k | HeikoS, any news? | 20:27 |
-!- shogun-notifier- [~irker@7nn.de] has joined #shogun | 20:32 | |
shogun-notifier- | shogun: Parijat Mazumdar :develop * 4390bbf / / (9 files): https://github.com/shogun-toolbox/shogun/commit/4390bbf357f63056256d61d308927e745d9e716c | 20:32 |
shogun-notifier- | shogun: minibatch kmeans added | 20:32 |
shogun-notifier- | shogun: Parijat Mazumdar :develop * b7fe2e6 / src/shogun/clustering/lKMeans.cpp: https://github.com/shogun-toolbox/shogun/commit/b7fe2e65ff06746dafb1506745e22a820879c268 | 20:32 |
shogun-notifier- | shogun: reverted to random algo | 20:32 |
shogun-notifier- | shogun: Fernando Iglesias :develop * 6930eed / / (9 files): https://github.com/shogun-toolbox/shogun/commit/6930eedec2d90c4e4b2029b0864fd118d77cf77f | 20:32 |
shogun-notifier- | shogun: Merge pull request #1827 from mazumdarparijat/fastkmeans | 20:33 |
shogun-notifier- | shogun: | 20:33 |
shogun-notifier- | shogun: mini-batch KMeans Added | 20:33 |
-!- thoralf [~thoralf@91-65-140-66-dynip.superkabel.de] has joined #shogun | 20:36 | |
thoralf | Hey! | 20:36 |
@iglesiasg | hello hello | 20:36 |
@sonney2k | hey thoralf | 20:37 |
thoralf | Did I miss something? :) | 20:37 |
@sonney2k | thoralf, heh | 20:38 |
lisitsyn | are we in that stamstststsich state? | 20:39 |
@sonney2k | lisitsyn, I am | 20:40 |
@sonney2k | lisitsyn, but I only have 20 minutes before being kicked out | 20:40 |
@sonney2k | internet here sucks big time | 20:40 |
lisitsyn | heh | 20:40 |
@HeikoS | sonney2k: hi! | 20:43 |
@HeikoS | lisitsyn: hi! | 20:43 |
@HeikoS | sonney2k: I have two projects, talking to a guy about a third one | 20:44 |
@HeikoS | I could do another one, but thats out of my time constraints ;) | 20:44 |
@sonney2k | HeikoS, so it seems we have enough projects ... | 20:45 |
@HeikoS | sonney2k: thats good! | 20:45 |
@HeikoS | I will add two descriptions soon end of weekish | 20:45 |
lisitsyn | HeikoS: hey | 20:45 |
-!- benibadman [~benibadma@port-92-206-75-57.dynamic.qsc.de] has joined #shogun | 20:46 | |
-!- benibadman [~benibadma@port-92-206-75-57.dynamic.qsc.de] has quit [Read error: Connection reset by peer] | 20:47 | |
-!- benibadman [~benibadma@port-92-206-75-57.dynamic.qsc.de] has joined #shogun | 20:47 | |
@sonney2k | HeikoS, I cannot start doing stuff before next week | 20:48 |
@sonney2k | HeikoS, internet sucks here so I am basically offline | 20:48 |
lisitsyn | sonney2k: where you at? ;) | 20:48 |
@sonney2k | HeikoS, btw regarding funding - we had a paid developer at some stage. So I know that money is not the key to a OSS project's success | 20:49 |
@sonney2k | lisitsyn, Oberwiesenthal | 20:49 |
lisitsyn | sonney2k: is it dirk who is a paid developer? | 20:50 |
lisitsyn | :P | 20:50 |
@HeikoS | sonney2k: ok no worries there is still time | 20:50 |
lisitsyn | ahh sonney2k having ski time? | 20:50 |
lisitsyn | I am zero-experienced but I tend to agree it is not about money, HeikoS sonney2k | 20:51 |
@HeikoS | sonney2k: funding would still be nice, but yeah, money is not the thing | 20:51 |
@HeikoS | but continuity :) | 20:51 |
@sonney2k | lisitsyn, indeed but he wasn't the only one | 20:51 |
@HeikoS | so if we could pay sergey, that would be cool | 20:51 |
@sonney2k | HeikoS, money != continuity | 20:52 |
@HeikoS | sonney2k: well up to some point it is | 20:52 |
@HeikoS | just hour wise | 20:52 |
lisitsyn | HeikoS: even in this case I could fail :D | 20:52 |
@sonney2k | yes for a few months | 20:52 |
shogun-buildbot | build #449 of FC19 - libshogun is complete: Failure [failed test] Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20libshogun/builds/449 blamelist: Parijat Mazumdar <mazumdarparijat@gmail.com> | 20:52 |
@sonney2k | but then it is just work | 20:52 |
@sonney2k | and you loose interest in doing it the way you did... | 20:52 |
@sonney2k | but of course there are exceptions | 20:52 |
@HeikoS | sonney2k: its worth a try | 20:53 |
@HeikoS | but yeah, as you said, gsoc first | 20:53 |
@sonney2k | lisitsyn, you are mr. chaos(tm) anywas :) | 20:53 |
lisitsyn | sonney2k: sure | 20:53 |
@sonney2k | HeikoS, yes indeed | 20:53 |
lisitsyn | sonney2k: isn't it the way everything should work? | 20:53 |
lisitsyn | chaotic! | 20:53 |
@sonney2k | GSoC gave use much more continuity :) | 20:53 |
lisitsyn | I have to admit I don't get continuity word here | 20:53 |
@sonney2k | lisitsyn, that is how our team works :D | 20:54 |
@sonney2k | all entropy maximizers! | 20:54 |
@HeikoS | haha | 20:54 |
lisitsyn | yeah sure I thought max ent is the way | 20:54 |
lisitsyn | :D | 20:54 |
@HeikoS | sonney2k: still would be good to have some less new-algorithm based person working on stuff | 20:54 |
@sonney2k | yeah we had that - it is boring and at some point you wish you would do sth else | 20:55 |
@sonney2k | but of course this helped | 20:55 |
@HeikoS | yeah one needs a good mix | 20:55 |
@sonney2k | but not regarding continuity | 20:55 |
@sonney2k | difficult | 20:55 |
@HeikoS | and change targets every now and then | 20:55 |
@HeikoS | I think we can only win by this :) | 20:55 |
@sonney2k | ML people don't like non-new-algorithm stuff | 20:55 |
@HeikoS | wiking does :) | 20:55 |
@HeikoS | sergey too | 20:55 |
@sonney2k | and hardcore coders get bored with lowlevel stuff | 20:55 |
lisitsyn | HeikoS: we are not ml people ;) | 20:56 |
@HeikoS | I dont see this so pessimistic and would see its good to try things, | 20:56 |
@sonney2k | HeikoS, one pressing thing is getting the money btw | 20:56 |
@HeikoS | every now and then | 20:56 |
@sonney2k | HeikoS, sure do it | 20:56 |
@HeikoS | sonney2k: yeah just reading your email | 20:56 |
@sonney2k | HeikoS, you just should not be 100% away | 20:56 |
@HeikoS | sonney2k: I need that money too, fincance is pressing here | 20:56 |
@HeikoS | sonney2k: I was 100% away from everything over the last months | 20:56 |
@HeikoS | but that wont happen anymore in the near/middle future | 20:57 |
@HeikoS | (I hope at least) | 20:57 |
@sonney2k | HeikoS, I know | 20:57 |
@HeikoS | kind of crazy period here | 20:57 |
@sonney2k | just after the summit | 20:57 |
lisitsyn | you must feel good now | 20:57 |
@sonney2k | pretty bad :( | 20:57 |
@HeikoS | did not do anything else than working and sleeping | 20:57 |
@HeikoS | I do feel free :) | 20:57 |
@sonney2k | btw guys we have this buildbot here showing very critical valgrind errors | 20:57 |
@sonney2k | http://buildbot.shogun-toolbox.org/builders/debian%20wheezy%20-%20memcheck/builds/165 | 20:57 |
lisitsyn | HeikoS: how did you manage to concentrate? :D | 20:57 |
@HeikoS | but now there is all the postponed stuff waiting ;) | 20:57 |
@HeikoS | lisitsyn: well focus on one thing rather than many | 20:57 |
@sonney2k | it is unfortunately green all the time | 20:57 |
lisitsyn | HeikoS: I constantly fail in this | 20:57 |
@HeikoS | sonney2k: whats the problem with that? | 20:58 |
-!- HeikoS [~heiko@nat-164-138.internal.eduroam.ucl.ac.uk] has left #shogun [] | 20:58 | |
-!- HeikoS [~heiko@nat-164-138.internal.eduroam.ucl.ac.uk] has joined #shogun | 20:58 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 20:58 | |
@sonney2k | HeikoS, it does valgrind on the tests | 20:58 |
@sonney2k | HeikoS, and many have uninted memory reads / errors | 20:58 |
-!- HeikoS [~heiko@nat-164-138.internal.eduroam.ucl.ac.uk] has left #shogun [] | 20:58 | |
-!- HeikoS [~heiko@nat-164-138.internal.eduroam.ucl.ac.uk] has joined #shogun | 20:58 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 20:58 | |
@sonney2k | HeikoS, and some of them fail | 20:58 |
@HeikoS | sonney2k: I see | 20:59 |
@sonney2k | on osx | 20:59 |
@HeikoS | bad bad | 20:59 |
lisitsyn | oh I think I have a new idea to propose | 20:59 |
@sonney2k | they will kick me out of this lounge here in 1 min... | 20:59 |
lisitsyn | dictionary learning | 20:59 |
lisitsyn | the problme is I am not an expert in that at all | 20:59 |
lisitsyn | is it a problem for any of you? | 20:59 |
@HeikoS | sonney2k: whats the most important stuff to discuss for now | 20:59 |
shogun-buildbot | build #450 of FC19 - libshogun is complete: Success [build successful] Build details are at http://buildbot.shogun-toolbox.org/builders/FC19%20-%20libshogun/builds/450 | 20:59 |
@sonney2k | HeikoS, well get the money from previous gsoc | 21:00 |
@sonney2k | get into this years gsoc | 21:00 |
@sonney2k | so proposal needs to be written | 21:00 |
@HeikoS | sonney2k: ok, so we dont need more projects? just proposals? | 21:00 |
@sonney2k | and ideas list in shape | 21:00 |
@HeikoS | if I add two, we have some things? | 21:00 |
@HeikoS | new mentors? | 21:00 |
@HeikoS | neural net? variational learning? | 21:00 |
@sonney2k | I asked many people and some of them said yes so we should be fine | 21:00 |
@sonney2k | we have to finish it though but one thing at a time | 21:01 |
@sonney2k | then I would say | 21:01 |
@sonney2k | get builbots to green | 21:01 |
@sonney2k | and criticals fixed | 21:01 |
@sonney2k | and then release... | 21:01 |
@HeikoS | sonney2k: did you submit an amount along with the pdf invoices you sent? | 21:01 |
@sonney2k | and then gsoc students Q/A as usual | 21:01 |
@sonney2k | ohh btw | 21:01 |
@sonney2k | HeikoS, I thought I CC'ed you?! | 21:02 |
@sonney2k | but he actually did | 21:02 |
@sonney2k | ohh an btw workshop - > before eurypython in berlin | 21:02 |
@sonney2k | Europython | 21:02 |
@wiking | mmm | 21:02 |
@HeikoS | sonney2k: phew load sof stuff | 21:02 |
@wiking | stamtisch :) | 21:02 |
@HeikoS | Ill try to do something little every day ;) | 21:02 |
@wiking | or wtf :> woah i've managed :> | 21:02 |
@sonney2k | HeikoS, sure | 21:03 |
@HeikoS | sonney2k: no you did not, do you still find the email? | 21:03 |
@sonney2k | HeikoS, the guy attached the bill in his email | 21:03 |
@sonney2k | HeikoS, so you have it too | 21:03 |
@HeikoS | sonney2k: I don tknow whether I have to put a number | 21:03 |
@HeikoS | or just the scans and they already have the sum | 21:03 |
@sonney2k | IIRC all he wanted is that we prepend some bill to google | 21:03 |
@HeikoS | wiking: hi there! how are things? | 21:03 |
@wiking | HeikoS: pfff | 21:03 |
@sonney2k | HeikoS, they have everything | 21:03 |
@wiking | HeikoS: but thanks for asking :) | 21:03 |
@sonney2k | HeikoS, wiking is in his final phase of phd | 21:04 |
@sonney2k | so closer to suicide as usual ;) | 21:04 |
lisitsyn | oh shit I should write my masters thesis | 21:04 |
lisitsyn | good you recalled | 21:04 |
@sonney2k | (every phd student knows what I am talking about) | 21:04 |
@HeikoS | oh | 21:04 |
lisitsyn | :D | 21:04 |
@HeikoS | wiking: haha ;) I share your pain | 21:04 |
@wiking | HeikoS: thnx | 21:04 |
@sonney2k | HeikoS, then I think I should try to get shogun properly into debian | 21:04 |
@sonney2k | we have sth in now | 21:04 |
@sonney2k | but no python etc | 21:04 |
@HeikoS | sonney2k: I agree | 21:05 |
@sonney2k | HeikoS, besser82 already gave us nice fedora packages | 21:06 |
@sonney2k | so we are in much better shape than the years before | 21:06 |
@sonney2k | HeikoS, would be nice if wiking could take care of osx stuff like brew / macports | 21:06 |
@sonney2k | and we need our binary packages on the website somewhere | 21:06 |
@sonney2k | alrighty | 21:06 |
@sonney2k | I am being kicked out | 21:07 |
@sonney2k | see you guys | 21:07 |
@wiking | cya | 21:07 |
lisitsyn | okay now people would notice my messages again :D | 21:07 |
@wiking | :D | 21:07 |
thoralf | lisitsyn ;) | 21:07 |
@wiking | lisitsyn: go and work on msc :) | 21:07 |
@wiking | so what's with the licensing on the end | 21:07 |
lisitsyn | wiking: I don't know what is it about | 21:07 |
@wiking | do we have a conclusion, or the conclusion is that we have no conclusion :) | 21:08 |
@wiking | okok i see | 21:08 |
lisitsyn | okay guys any suggestions for masters thesis? :D | 21:08 |
@HeikoS | sonney2k: see you! | 21:08 |
lisitsyn | I have 3 months essentially :D | 21:08 |
@wiking | well then we should really think that if we get accepted for gsoc this year | 21:08 |
@HeikoS | man deep mind just got bought by google | 21:08 |
@wiking | that maybe the students should send in their stuff with different license | 21:08 |
@wiking | HeikoS: yes | 21:08 |
@HeikoS | another scary bit in their puzzle | 21:08 |
lisitsyn | wiking: I vote for BSD/MIT | 21:08 |
lisitsyn | for new code | 21:09 |
@wiking | no it's scary how much they paid for it | 21:09 |
lisitsyn | ;) | 21:09 |
@HeikoS | wiking: deep mind is next door here and most people from the institute go there after phd | 21:09 |
@wiking | lisitsyn: could go with it | 21:09 |
@wiking | HeikoS: hehe thought so | 21:09 |
@HeikoS | pretty cool caompany, would have wanted to go there | 21:09 |
@HeikoS | but now....aaargh | 21:09 |
@HeikoS | wiking: whats with this bsd change? | 21:09 |
@HeikoS | wiking: how much work is that? | 21:09 |
-!- zxtx [~zv@129-79-241-148.dhcp-bl.indiana.edu] has joined #shogun | 21:09 | |
besser82 | HeikoS: a f*cking bunch | 21:09 |
@wiking | HeikoS: well it's more waiting than | 21:09 |
lisitsyn | hah | 21:09 |
@HeikoS | if we could do this kind of on the fly every now and then continously that would be great | 21:09 |
@wiking | work actually | 21:09 |
@HeikoS | wiking: yeah thats what I mean | 21:09 |
@HeikoS | and we can ask all code from now to be bsd | 21:10 |
@HeikoS | or mit | 21:10 |
@HeikoS | or whatever | 21:10 |
@wiking | that's not an issue | 21:10 |
besser82 | HeikoS, lisitsyn, wiking, adrin, iglesiasg, thoralf: Hey Folks!!! | 21:10 |
@wiking | the past code and their writers has to be contacted | 21:10 |
@wiking | yoyo | 21:10 |
@HeikoS | besser82: hi! :) | 21:10 |
thoralf | Hey Bj?rn | 21:10 |
lisitsyn | is there anyone else (but HeikoS) who thinks we should kinda split up to a few smaller libraries? | 21:10 |
@HeikoS | wiking: how much automatic would that be | 21:10 |
@HeikoS | lisitsyn: is that a gsoc project? | 21:11 |
@wiking | HeikoS: well i've done some basic stats by git | 21:11 |
@wiking | it's doable | 21:11 |
lisitsyn | this should solve licensing problem a bit | 21:11 |
lisitsyn | I mean we just replace parts | 21:11 |
@HeikoS | wiking: ok, so maybe I will write an email and put it to disciussion this week | 21:11 |
lisitsyn | do we need massive -> bsd conversion? | 21:11 |
@wiking | i mean the core developers can right away assign to the right license | 21:11 |
@wiking | the problem is with contributors | 21:11 |
@wiking | who are gone | 21:11 |
@HeikoS | lisitsyn: yeah I would just try things first and then see how it goes | 21:11 |
@HeikoS | wiking: yep | 21:11 |
@wiking | but in worst case | 21:11 |
@HeikoS | would be good to have a map of things | 21:12 |
@wiking | we can do a rewrite of that code | 21:12 |
@wiking | or let's say this is a good way to throw out code | 21:12 |
@wiking | HeikoS: mindmap? :) | 21:12 |
lisitsyn | wiking: what's with berlin universitaet copyrights? | 21:12 |
@HeikoS | wiking: yeah maybe later, I would first change everything thats possible and start changing new stuff | 21:12 |
@wiking | lisitsyn: mmm has to be followed up | 21:12 |
@HeikoS | and then discuss what to do with the rest | 21:12 |
@HeikoS | but guys, what about gsoc, whats your status on that? project ideas? ideas page, proposals? application, who manages? | 21:13 |
lisitsyn | wiking: is it enough to have author will to change anything? | 21:13 |
@wiking | lisitsyn: yes | 21:13 |
lisitsyn | u sure? | 21:13 |
@wiking | lisitsyn: but of course if there's an institution connected | 21:13 |
@wiking | lisitsyn: then u have to get the institution to sign of the change as well | 21:13 |
lisitsyn | haha | 21:13 |
lisitsyn | then gsoc 2011 work is kind of lost | 21:13 |
@wiking | HeikoS: i've promised to write the proposal | 21:13 |
@wiking | lisitsyn: y? | 21:13 |
@HeikoS | wiking: on what? | 21:13 |
besser82 | lisitsyn: several smaller libs / plugin-stuff 4 sure! | 21:13 |
@wiking | HeikoS: for gsoc... and put in shape the whole project ideas | 21:14 |
lisitsyn | besser82: good to have people on your side :D | 21:14 |
@wiking | HeikoS: will do this till wednesday | 21:14 |
besser82 | HeikoS: me is mentoring SVM^bright && some cmake-based stuff | 21:14 |
@wiking | lisitsyn: what's the problem with 2011 contribs? | 21:14 |
@HeikoS | besser82: cool! | 21:14 |
lisitsyn | wiking: I mean they are all signed as berlin stuff | 21:14 |
lisitsyn | as soeren was in TU at the moment | 21:14 |
@wiking | lisitsyn: well we can get somebody in berlin to sign it off | 21:14 |
@wiking | lisitsyn: yeah but there's alex and some other guys still there | 21:15 |
lisitsyn | ah okay if it works I don't mind | 21:15 |
-!- benibadman [~benibadma@port-92-206-75-57.dynamic.qsc.de] has quit [Ping timeout: 245 seconds] | 21:15 | |
@wiking | so i dont really think that we cannot get somebody in TU to sign of a piece of paper | 21:15 |
lisitsyn | I have absolutely no clue how this works | 21:15 |
@wiking | lisitsyn: i've checked up | 21:15 |
lisitsyn | cool | 21:15 |
lisitsyn | then we are ok | 21:15 |
lisitsyn | :) | 21:15 |
@wiking | we need basically a statement | 21:16 |
@wiking | from the parties that they are ok with the licensing change | 21:16 |
@wiking | that's all | 21:16 |
@HeikoS | lisitsyn, wiking we can just ask and see then | 21:16 |
@wiking | it's not rocket science | 21:16 |
@wiking | ok | 21:16 |
@wiking | so fucking let's go with BSD 2 liners | 21:16 |
@HeikoS | naywhayare: are you interested in having another project on this inter project comparison stuff? | 21:16 |
@HeikoS | wiking: yeah I agree on that | 21:16 |
@wiking | ok | 21:17 |
@wiking | i'll scetch up the statement for that | 21:17 |
@wiking | that we can send around to license holders in shogun | 21:17 |
@HeikoS | wiking: cool, maybe use the vlc stuff as inspiration | 21:17 |
@wiking | and then we are good | 21:17 |
@wiking | HeikoS: indeed | 21:17 |
@wiking | HeikoS: get funding :) | 21:17 |
@HeikoS | wiking: important part would be to maintain answers decisions and what parts are missing | 21:17 |
@wiking | if docker got 15M usd... | 21:17 |
@wiking | fuck man that pissed me off :) | 21:17 |
@HeikoS | wiking: I know, | 21:17 |
shogun-buildbot | build #45 of deb4 - python3 is complete: Failure [failed test python modular] Build details are at http://buildbot.shogun-toolbox.org/builders/deb4%20-%20python3/builds/45 blamelist: Parijat Mazumdar <mazumdarparijat@gmail.com> | 21:17 |
@HeikoS | well we can do that, just need to work harder ;)( | 21:17 |
@wiking | yeah as usual | 21:18 |
@HeikoS | wiking: we are growing, so patiance and discipline will solve this ;) | 21:18 |
lisitsyn | german | 21:18 |
@HeikoS | last year, nobody was even talking about funding | 21:18 |
@HeikoS | and the year before we did not even have unit tests | 21:18 |
@wiking | HeikoS: heheh yeah sure | 21:18 |
@HeikoS | notebooks are also new | 21:18 |
@HeikoS | so its going well! | 21:18 |
lisitsyn | HeikoS: DAS DISCIPLINE | 21:18 |
lisitsyn | ;) | 21:18 |
@wiking | lisitsyn: :D | 21:18 |
-!- benibadman [~benibadma@port-92-206-111-65.dynamic.qsc.de] has joined #shogun | 21:19 | |
@wiking | btw i have to load of some cash to some fundations | 21:19 |
@wiking | so plz do something with shogun asap | 21:19 |
@wiking | and then i'll transfer some money | 21:19 |
@HeikoS | wiking: oh | 21:19 |
@HeikoS | ok | 21:19 |
@HeikoS | yeah everything formal should be ready | 21:19 |
@wiking | HeikoS: not too much money | 21:19 |
@HeikoS | now we need to register | 21:19 |
@wiking | but still | 21:19 |
@wiking | shoudl be enough as a started | 21:19 |
@wiking | *starter | 21:19 |
lisitsyn | wiking: you got a money you can't spend? ;) | 21:20 |
@wiking | lisitsyn: nono | 21:20 |
@wiking | lisitsyn: it's money that i have to spend for associations/fundation, i.e. nonprofit | 21:20 |
@wiking | otherwise it's lost money | 21:20 |
@HeikoS | wiking: wherefrom? | 21:20 |
lisitsyn | heh yeah interesting ;) | 21:20 |
@wiking | because then the state will eat that money | 21:20 |
@wiking | so i rather give that money for something i know... | 21:20 |
@wiking | HeikoS: have a small company that has to do taxes | 21:21 |
@HeikoS | wiking: I see | 21:21 |
@HeikoS | very cool! | 21:21 |
@wiking | HeikoS: and i can assign freely 1% of my tax to nonprofit shiatz | 21:21 |
@HeikoS | hopefully next month we are set | 21:21 |
@wiking | as said if i dont assign it to somebody | 21:21 |
@wiking | then it's just lost in the big state budget | 21:21 |
@wiking | anyhow | 21:22 |
@wiking | let's hope we are still in time | 21:22 |
@wiking | i've talked with vojtech | 21:22 |
@wiking | about some qp project | 21:22 |
@wiking | it'll be a huge project | 21:22 |
@wiking | i dont know if anybody will be up for that task | 21:22 |
@HeikoS | ok | 21:22 |
@HeikoS | that would be cool | 21:22 |
@wiking | btw i could do gsoc this year (apparently i still have student status :DDD | 21:22 |
@HeikoS | haha :) | 21:23 |
@HeikoS | you should rather mentor two students ;) | 21:23 |
@wiking | yeah | 21:23 |
@wiking | figured | 21:23 |
@wiking | ok so lisitsyn r u going to lead the modularization ? :) | 21:24 |
@wiking | i mean if we get into gsoc | 21:24 |
lisitsyn | wiking: yeah I think it would work for me | 21:24 |
@wiking | cool | 21:24 |
@wiking | that should have a prio among the slots | 21:24 |
@wiking | imho | 21:25 |
@wiking | if there's a good student up for that task | 21:25 |
lisitsyn | wiking: I realized I am rather a gay^W^W a software developer than machine learning guy | 21:25 |
lisitsyn | :D | 21:25 |
@wiking | lisitsyn: :D | 21:25 |
@wiking | btw the whole story with dptrs would be still required even if we do modularization? | 21:25 |
lisitsyn | wiking: I think d-ptr stuff is useful anyways | 21:26 |
besser82 | wiking: dtpr would be good in some parts, yes ;) | 21:26 |
besser82 | lisitsyn: :D | 21:26 |
@wiking | i guess we would still benefit a lot from it | 21:26 |
lisitsyn | I believe once we get better compile time | 21:26 |
@wiking | for the swig interfaces | 21:26 |
@wiking | or? | 21:26 |
lisitsyn | wiking: don't know | 21:26 |
lisitsyn | we will see | 21:26 |
besser82 | wiking: yes, swig would be a cleaner solution then | 21:27 |
@wiking | lisitsyn: yeah i think the generated cxx would be heaps smaller | 21:27 |
@wiking | HeikoS: wha't swith that view thing atm? | 21:27 |
besser82 | wiking: and generation of them cxx wouldn't take 1.5GBytes each anymore :-P | 21:27 |
-!- dsockwell [~twgs@199.167.199.97] has quit [Ping timeout: 272 seconds] | 21:28 | |
@wiking | :) | 21:28 |
lisitsyn | oh guys guys | 21:28 |
@HeikoS | wiking: the guy did not finish it unfortunately | 21:28 |
lisitsyn | I have an idea | 21:28 |
@wiking | HeikoS: what's missing? | 21:28 |
@HeikoS | wiking: the solution was nice though | 21:28 |
@HeikoS | should not be hard to finish it | 21:28 |
lisitsyn | lets put whole shogun code to one .h | 21:28 |
@wiking | HeikoS: if u tell me what's missing i could give it a go | 21:28 |
@HeikoS | wiking: some playing around with it to see it works and fix copy constructors of some feature classes | 21:28 |
@wiking | lisitsyn: !!! | 21:28 |
lisitsyn | what's your bet on required RAM? | 21:28 |
@wiking | HeikoS: indeed | 21:28 |
@HeikoS | wiking: have a look at the current PR of the guy | 21:28 |
@wiking | HeikoS: will do | 21:28 |
@HeikoS | wiking: and then change the x-validation class to use it as a test case (and write unit tests :) | 21:28 |
@HeikoS | and then it should be ready to be used by others | 21:28 |
lisitsyn | I think 30gb should be the number :D | 21:28 |
@HeikoS | and if you dont change the copy-by-reference objects, it is threadsafe | 21:29 |
@wiking | HeikoS: ok so not muc :) | 21:29 |
@wiking | HeikoS: well i might just skip xvalidation | 21:29 |
@wiking | HeikoS: but test it in other places :PPP | 21:29 |
@wiking | HeikoS: openmp created shiatz | 21:29 |
-!- dsockwell [~twgs@199.167.199.97] has joined #shogun | 21:30 | |
@HeikoS | wiking: I think it would be good to test it with xval since its a good test case and there are already things relying on it so we can see whather stuff breaks | 21:30 |
@HeikoS | wiking, lisitsyn I have to go now. See you tomorrow! | 21:31 |
@wiking | HeikoS: cyaz | 21:31 |
-!- HeikoS [~heiko@nat-164-138.internal.eduroam.ucl.ac.uk] has quit [Quit: Leaving.] | 21:31 | |
lisitsyn | wiking: I want to have dictionary learning in shogun too | 21:34 |
lisitsyn | may be I could mentor it somehow | 21:34 |
lisitsyn | at least some ksvd | 21:34 |
@wiking | mmm | 21:34 |
@wiking | lisitsyn: we can steal the code :) | 21:34 |
lisitsyn | yeah sure | 21:34 |
lisitsyn | everything is already implemented | 21:34 |
@wiking | 'steal' = opensource | 21:34 |
@wiking | yeah so that shouldn't be a biggie | 21:35 |
@wiking | i mean we can put together such a slot as well | 21:35 |
@wiking | where u dont need too much of a haxor skill | 21:35 |
lisitsyn | yeah | 21:35 |
@wiking | rather just do fast integration of some crucial basic algo | 21:35 |
lisitsyn | cause plugins is a hardcore thing | 21:35 |
@wiking | that are atm missing from shogun | 21:35 |
@wiking | yeah that's like pretty fucking crayz | 21:35 |
lisitsyn | we ain't gonna have a student for that | 21:36 |
lisitsyn | :D | 21:36 |
@wiking | hehehe | 21:36 |
@wiking | i can come as a student as i told :DDD | 21:36 |
@wiking | but imo i wont have time | 21:36 |
@wiking | :< | 21:36 |
@wiking | thoralf: i see it's rather a longer fix | 21:37 |
@wiking | thoralf: there were a lot of code redundancy in the lib* implementation | 21:37 |
@wiking | and as it's a matlab -> c port | 21:38 |
@wiking | i wouldn't be too surprised of such 'little' bug | 21:38 |
@wiking | s | 21:38 |
thoralf | wiking: I hope I finally convinced Michal ;) | 21:38 |
-!- travis-ci [~travis-ci@ec2-50-19-5-30.compute-1.amazonaws.com] has joined #shogun | 21:39 | |
travis-ci | [travis-ci] it's Fernando Iglesias's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: http://travis-ci.org/shogun-toolbox/shogun/builds/17718960 | 21:39 |
-!- travis-ci [~travis-ci@ec2-50-19-5-30.compute-1.amazonaws.com] has left #shogun [] | 21:39 | |
thoralf | wiking: Let me look to the other classes if they have similar issues... | 21:39 |
@wiking | thoralf: lib*bmrm are quite having the same codebase | 21:40 |
@wiking | they only have minor differences | 21:40 |
thoralf | Hmm, handling of beta is quite different in libp3bm and libncbm. | 21:46 |
thoralf | But libppbm.cpp seems to suffer the same issue. | 21:47 |
thoralf | line 265 | 21:47 |
naywhayare | HeikoS: sure; we have the framework in place and it could definitely be improved | 21:49 |
naywhayare | oh, he's not even here. I should check first ... | 21:49 |
thoralf | wiking: Thanks for pointing that out. | 21:51 |
@wiking | naywhayare: which/what fw? | 21:54 |
@wiking | thoralf: nw... in the past i've hacked a bit that code | 21:55 |
@wiking | to make it a bit more shogun | 21:55 |
@wiking | ready | 21:55 |
@wiking | :P | 21:55 |
naywhayare | wiking: last year as a GSoC project mlpack had a student who made an automatic benchmarking system | 22:02 |
naywhayare | I met HeikoS at the mentor summit and we talked about how it would be useful for both of our libraries -- or really all ML libraries in general | 22:02 |
naywhayare | here's the last run of benchmarks we had... http://www.mlpack.org/benchmark.html | 22:03 |
naywhayare | anyway, it needs better documentation, better details, probably a better interface | 22:03 |
naywhayare | I think HeikoS was interested in adapting the system so it could be used by shogun | 22:03 |
@wiking | oh cool | 22:03 |
@wiking | naywhayare: is it hardwired with mlpack ? | 22:03 |
naywhayare | I also need to rebuild the benchmarks for your latest release but a paper deadline is on the 31st so it won't be done until after that... | 22:03 |
naywhayare | no, it's library-agnostic | 22:03 |
@wiking | great | 22:03 |
naywhayare | what the framework really is is just a big python script that runs things and compares how long they take | 22:04 |
naywhayare | it has a configuration file that defines what to run, what datasets to use, and how to run it | 22:04 |
naywhayare | and then it has little python scripts individually written for each library's implementation of each method | 22:04 |
-!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has quit [Quit: Leaving] | 22:05 | |
@wiking | naywhayare: and python runs like a binary or if there's a python interface u use that? | 22:05 |
@wiking | naywhayare: cool it seems there'a lot of stuff we need to fix :P | 22:07 |
@wiking | in shogun | 22:07 |
@wiking | i'm just checking GMM results | 22:07 |
@wiking | naywhayare: do u have somewhere the code released? | 22:08 |
naywhayare | wiking: yeah, here's the repo link... | 22:08 |
naywhayare | http://svn.cc.gatech.edu/fastlab/mlpack/conf/jenkins-conf/benchmark/ | 22:08 |
naywhayare | it wasn't really part of mlpack so we put it in our configuration directory for our build server | 22:08 |
naywhayare | and we have it all integrated with the build server so that it's easy to run tests | 22:08 |
naywhayare | here's some further documentation | 22:09 |
naywhayare | http://www.mlpack.org/trac/wiki/AutomaticBenchmark | 22:09 |
@wiking | ok cool | 22:09 |
@wiking | i'll try to run this myself | 22:09 |
naywhayare | like I said it's not completely polished | 22:10 |
naywhayare | we did our best to write the scripts for other libraries to use them as well as possible | 22:10 |
@wiking | no worries | 22:10 |
thoralf | wiking: libppbm has the same issue - just proven it in a unit test. | 22:10 |
@wiking | thoralf: \o/ | 22:10 |
naywhayare | the cover tree is not used for shogun k-nearest-neighbors because john langford's code has a bug he has yet to fix and it returns incorrect results (I think I mentioned this in here a while back; he still hasn't fixed it) | 22:11 |
@wiking | naywhayare: maybe we could start a separate repo for this on github? | 22:11 |
naywhayare | also, when you see '-' instead of 'failure' for the results, it means the test wasn't run for that particular dataset | 22:11 |
@wiking | naywhayare: and then we can get travis to do all the testing? :P | 22:11 |
naywhayare | that's just a configuration file issue and I'm in the process of fixing them little by little... | 22:11 |
@wiking | naywhayare: how long did it take for u to run the benchmark | 22:12 |
naywhayare | heh. the mlpack one takes two weeks | 22:12 |
naywhayare | the other libraries tend to take 3-5 days | 22:12 |
@wiking | oh woah :D | 22:12 |
naywhayare | with mlpack we do memory profiling too | 22:12 |
@wiking | naywhayare: with? | 22:12 |
naywhayare | massif/valgrind | 22:12 |
@wiking | naywhayare: do u use some fancy malloc library or the base system? | 22:12 |
naywhayare | just valgrind, I'm not 100% sure on the details | 22:13 |
@wiking | ok | 22:13 |
naywhayare | if you click the rightmost button for a method, then you can open up the 'memory results' | 22:13 |
naywhayare | and it shows you memory usage for mlpack during the run of the algorithm | 22:13 |
@wiking | i see | 22:13 |
@wiking | cool | 22:13 |
@wiking | would be great to have that for all the libraries of coursE :P | 22:13 |
naywhayare | yeah; we did what we could, but time is so little that it's really hard to keep the effort going | 22:14 |
@wiking | but then again that's future | 22:14 |
naywhayare | I have enough trouble just keeping up to date with the five or six libraries we do test | 22:14 |
@wiking | naywhayare: totally understand | 22:14 |
naywhayare | because every time a new release happens I have to recompile it on my test system, then ensure the python scripts still work | 22:14 |
naywhayare | ...then wait 3-5 days for the results to be done :) | 22:14 |
@wiking | naywhayare: do u think we could get a separate repo for this | 22:14 |
@wiking | ? | 22:14 |
@wiking | so that anybody can contribute easily | 22:15 |
naywhayare | yeah, I mean, it's open-source. =) | 22:15 |
@wiking | i'm sure ppl here in shogun would be keen to contribute into it | 22:15 |
naywhayare | ideally the student should own the github repo, since it was his work and he knows the most about it | 22:15 |
@wiking | naywhayare: yeah no i mean that we could say that we create an official repo at github for it | 22:15 |
naywhayare | definitely | 22:15 |
@wiking | and then jsut have collaborators for the same repo | 22:15 |
@wiking | and then PRs can be sent etc | 22:15 |
@wiking | naywhayare: r u still in touch with the student? | 22:16 |
naywhayare | I'm getting in touch with him about it (his irc nick is 'marcus_zoq' and he hangs out in #mlpack or you can pm him); I'll let you know what he says | 22:16 |
naywhayare | yeah, we still talk regularly, which is good | 22:16 |
@wiking | great | 22:16 |
@wiking | could u please ask him if he would do this? | 22:16 |
shogun-buildbot | build #46 of deb4 - python3 is complete: Failure [failed test python modular] Build details are at http://buildbot.shogun-toolbox.org/builders/deb4%20-%20python3/builds/46 blamelist: Fernando Iglesias <fernando.iglesiasg@gmail.com> | 22:16 |
naywhayare | I just did, don't worry =) | 22:16 |
@wiking | cool great! | 22:16 |
naywhayare | really github is a better place anyway than hidden somewhere inside our jenkins configuration | 22:16 |
@wiking | heheh yeah | 22:16 |
@wiking | just with a small link to that page in README.md | 22:17 |
@wiking | and that's all | 22:17 |
@wiking | after that issues and PRs will fly anyways | 22:17 |
naywhayare | we submitted the project to the MLOSS workshop but the reviewers bashed it because they were more interested in comparing the accuracy of algorithms instead of the runtime | 22:17 |
naywhayare | whereas for the most part we try to ensure that we are benchmarking the exact same algorithms against each other | 22:17 |
@wiking | hehehh yeah man | 22:17 |
@wiking | i mean reviewers are like a joke sometimes | 22:17 |
@wiking | i mean there's like 8 out of 10 papers | 22:18 |
naywhayare | nah, it wasn't a bad review, his points were valid, but sort of missed the point of what we were doing a bit | 22:18 |
@wiking | where they compare algos | 22:18 |
@wiking | but then u know that they fine tuned the last bit of their algo in the paper | 22:18 |
@wiking | while the ones they used for comparison | 22:18 |
@wiking | just used as is (default parameters etc) | 22:18 |
naywhayare | yeah, naturally. it's really difficult to get a clear picture like that | 22:18 |
@wiking | so really a thing like this would be great to have | 22:18 |
@wiking | jsut would be great to automate a lot of part of it | 22:19 |
@wiking | i mean the test generation | 22:19 |
naywhayare | so you can use kernel PCA for denoising and alex smola's big paper on it from 1999 has some awesome results... but that paper doesn't mention that when you do image denoising with kpca you still end up with outliers that have incredibly poor quality | 22:19 |
@wiking | :P | 22:19 |
@wiking | alex alex | 22:19 |
@wiking | i wonder where's he now | 22:19 |
naywhayare | that was a thing I learned this weekend. overall, the average reconstructed image quality is great! but there are just some terrible outliers sometimes. I wonder if anyone's investigated that | 22:19 |
@wiking | used to be in canberra | 22:19 |
naywhayare | either CMU or Google or Yahoo | 22:19 |
naywhayare | I met him at ICML... his poster was next to mine | 22:20 |
naywhayare | I thought that would be great for the publicity of my poster | 22:20 |
naywhayare | but actually everyone just talked to him and not me :( | 22:20 |
@wiking | heheheheheh | 22:20 |
@wiking | yeah usual | 22:20 |
@wiking | sorry to hear that though | 22:20 |
@wiking | btw how to u store the results? | 22:20 |
@wiking | i mean is it stored in a db or something? | 22:20 |
naywhayare | yeah, sqlite db | 22:20 |
@wiking | ok cool | 22:20 |
@wiking | so basically on the end u can generate historical data as well | 22:21 |
@wiking | right? | 22:21 |
naywhayare | yeah, although I don't think we have the support to do that well | 22:21 |
naywhayare | not right now at least | 22:21 |
@wiking | yeah sure | 22:21 |
naywhayare | but it can be done; there's just nothing in there that will generate those graphs | 22:21 |
@wiking | but if it's stored with a timestamp | 22:21 |
@wiking | it's really just up to a php/python script to do that | 22:21 |
naywhayare | http://big.cc.gt.atl.ga.us/ there's our build server. if you open the 'z benchmark' category then you can see each job that we run | 22:21 |
naywhayare | so when a new release happens, I update it on the system that will be running on the benchmarks (compile with -O3 and whatever), then click 'start job' on jenkins | 22:22 |
@wiking | cool | 22:22 |
@wiking | great stuff | 22:22 |
@wiking | i think there's a lot of great stuff that can be done with this | 22:22 |
@wiking | i mean on the end if u think about it... if this benchmark gets more wide | 22:23 |
@wiking | library releasers will themselves generate the output | 22:23 |
@wiking | of their new release | 22:23 |
naywhayare | yeah, I agree | 22:23 |
naywhayare | but again it gets really difficult to compare algorithms sometimes | 22:23 |
naywhayare | I can compare nearest-neighbor pretty easily because it's exact, but... do I compare it with kd-trees, or with the library's default trees? | 22:24 |
@wiking | well of course | 22:24 |
naywhayare | is the goal of the benchmark to show what a typical user might actually write in code, or to show the actual capabilities of the library? it gets difficult to answer the question cleanly | 22:24 |
@wiking | but there's always going to be problems like that | 22:24 |
naywhayare | because it may be that while I can use mlpack's AllkNN object, it's 5x as fast if I write a far more complex use of some other class | 22:24 |
@wiking | but we dont want to solve everything with this benchmark | 22:25 |
@wiking | rather to show at least basic stuff | 22:25 |
@wiking | and compare them | 22:25 |
naywhayare | yeah. I see those types of problems as a potential source of arguments that I really want to avoid. I don't want to hurt anyone's feelings :) | 22:25 |
@wiking | nono i mean i think everybody understands this | 22:25 |
naywhayare | there definitely needs to be a lot of clarification on what exactly each benchmark is doing, because it's unclear right now (for instance it just says "ALLKNN") | 22:25 |
@wiking | but this is good for everybody both users and library developers | 22:25 |
@wiking | at least get the bigger picture | 22:25 |
naywhayare | it's super useful for an internal mlpack development tool, though | 22:25 |
naywhayare | true | 22:26 |
@wiking | so it's really good ... only that we need to get some more people on board | 22:26 |
@wiking | but i think that will not be a problem | 22:26 |
naywhayare | yeah. I haven't had the time that I really need to put into this to make it as nice as I'd like it to be | 22:27 |
naywhayare | anyway a few good GSoC students can get a lot done :) | 22:27 |
@wiking | well dont try to do all yourselves :) | 22:27 |
@wiking | just put together a repo for it | 22:27 |
@wiking | and fire up the issues | 22:27 |
naywhayare | yeah | 22:27 |
@wiking | i think things will get started | 22:27 |
naywhayare | my hope is that Marcus will want to run that | 22:27 |
naywhayare | after all, it's his work :) | 22:28 |
@wiking | heheh yeah | 22:28 |
@wiking | naywhayare: the good thing would be to generate a standard API where u can submit your results | 22:28 |
@wiking | via http | 22:28 |
naywhayare | I haven't yet approached the scikit-learn guys. I told myself I'd do that when I had time but it looks like I'll never actually have time ever again | 22:28 |
naywhayare | yeah, but the benchmarks aren't useful if they're not run on the same system | 22:28 |
@wiking | yeah | 22:29 |
naywhayare | which is why we limit our runs to one specific system | 22:29 |
@wiking | but one can run the whole test | 22:29 |
@wiking | on another system | 22:29 |
naywhayare | ah, yeah | 22:29 |
@wiking | that's useful as well | 22:29 |
@wiking | then u see what's the difference between machine A and machine B | 22:29 |
@wiking | there's tons of things that could be done with this | 22:30 |
@wiking | so let's get marcus on board :) | 22:30 |
naywhayare | yeah. making a good interface to be able to actually parse all that data is the really tricky part | 22:30 |
naywhayare | or, I'm bad at interfaces :) | 22:30 |
@wiking | hehehe | 22:30 |
@wiking | i mean honestly it'll be a great thing if this gets picked up around 2015 | 22:31 |
@wiking | that's my optimistic opinion :P | 22:31 |
naywhayare | :) | 22:31 |
@wiking | there's still a lot to work on, but i believe everybody would be interested in this if it would get a little bit more attention | 22:31 |
@wiking | :P | 22:31 |
naywhayare | sonney2k pointed out that they tried a similar thing some years ago with http://www.mldata.org but did not see widespread adoption | 22:32 |
@wiking | one never knows when things go haywire :P | 22:33 |
-!- benibadman [~benibadma@port-92-206-111-65.dynamic.qsc.de] has quit [Remote host closed the connection] | 22:35 | |
@wiking | "US reaches deal with Internet companies on requests to reveal demands for customer information" | 22:35 |
-!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has joined #shogun | 22:35 | |
-!- mode/#shogun [+o iglesiasg] by ChanServ | 22:35 | |
@wiking | 2 more years and there's gonna b like usa intranet and rest of the world internet if it continues like this | 22:36 |
naywhayare | I'd like to hope Edward's Snowden's revelations will cause some kind of useful reforms. but I'm cynical and I doubt it... | 22:38 |
@wiking | ehehhe | 22:38 |
@wiking | naywhayare: yeah like with obama and change, right? :D | 22:38 |
@wiking | in retrospect that was like one of the biggest scam of the states :) | 22:39 |
naywhayare | in a country as big as ours (or really any big country) it's very easy to be cynical about politics because a single person's opinions can't really effect any change. which is very frustrating | 22:40 |
@wiking | ehheheh | 22:40 |
naywhayare | on a very local level, though, like a neighborhood / city level, I've had good interactions with local politicians | 22:40 |
@wiking | well things are everywhere the same... | 22:41 |
@wiking | i bet there's no one single person here who could tell the opposite | 22:41 |
naywhayare | hah, probably | 22:41 |
@wiking | whenever i meet somebody from another country | 22:42 |
@wiking | there's one single thing that everybody has in common | 22:42 |
@wiking | things are not going well in politics :) | 22:42 |
naywhayare | hah | 22:42 |
@wiking | i remember meeting this guy from brasil tell me stories of his own politicans, and in the meanwhile i was like 'yeah i know the story, they did the same...' | 22:43 |
@wiking | so it's all the same imo | 22:43 |
@wiking | i bet iglesiasg could share some very positive stuff from the royal family of spain :PPPPP | 22:44 |
naywhayare | I was in spain recently, I thought it was really nice :) | 22:44 |
@wiking | heheheheh | 22:44 |
naywhayare | but I didn't deal with the politics at all... :) | 22:44 |
@wiking | yeah that's the rule i always follow when i'm abroad | 22:44 |
@wiking | ignore local politics | 22:44 |
@wiking | things are just heaps better then :) | 22:44 |
@wiking | so lately i follow that rule for my own country's politics... so far so good : | 22:45 |
@wiking | :) | 22:45 |
naywhayare | heh | 22:45 |
@wiking | man i fucking hate hadoop sometimes :) | 22:46 |
@wiking | the moment when your 4th job out of 5 jobs dies... | 22:46 |
@wiking | after 12 hours :) | 22:46 |
@wiking | of running time | 22:46 |
naywhayare | oops | 22:47 |
@wiking | with a big fat exception | 22:47 |
@wiking | that basically doesn't really say anything | 22:47 |
@wiking | even an OOM would be more meaningful | 22:47 |
-!- HeikoS [~heiko@nat-187-115.internal.eduroam.ucl.ac.uk] has joined #shogun | 23:12 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 23:12 | |
-!- HeikoS [~heiko@nat-187-115.internal.eduroam.ucl.ac.uk] has quit [Client Quit] | 23:14 | |
thoralf | wiking: Err. I found a funny bug, that actually has no effect, but maybe you've got an idea how to fix it: | 23:23 |
thoralf | every lib*bm defines a BufSize | 23:23 |
thoralf | When calling any solver, the "local" version is used. | 23:24 |
thoralf | But when calling libbmrm::clean_icp() and adding an assertion there, the libbmrm-version is used, no matter from which solver it's been called. | 23:24 |
@wiking | mmm | 23:31 |
@wiking | that's most probably a bug introduced by me | 23:31 |
@wiking | because before that the code was just copy pasted among the 3 .cpp | 23:31 |
thoralf | lol | 23:31 |
@wiking | so i wanted to remove some redundancy | 23:31 |
@wiking | but then i guess i've made there a mistake | 23:32 |
thoralf | Actually is has no effect, but it's irritating as soon as you use BufSize from somewhere else. ;) | 23:32 |
@wiking | sorry for that | 23:32 |
-!- shogun-notifier- [~irker@7nn.de] has quit [Quit: transmission timeout] | 23:33 | |
thoralf | Can it be fixed by using the right qualifiers for BufSize? Something like extern uint BufSize instead of static? | 23:34 |
@wiking | mmm nasdaq is down 3 days in a row :< | 23:34 |
@wiking | thoralf: yeah that should work | 23:34 |
thoralf | I'll check. | 23:35 |
thoralf | wiking: Err. NASDAQ? What happened? | 23:38 |
thoralf | Didn't read news for days... | 23:39 |
@wiking | thoralf: mmm well since friday some countries' market is under siege | 23:39 |
@wiking | so it's all down now :) | 23:39 |
@wiking | but mainly it's FED's fault.... they are starting to pull the plug :) | 23:40 |
@wiking | tapering started basically :) | 23:41 |
-!- iglesiasg [~iglesiasg@524AE0A7.cm-4-3d.dynamic.ziggo.nl] has quit [Quit: Leaving] | 23:50 | |
--- Log closed Tue Jan 28 00:00:13 2014 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!